HyperAIHyperAI

Command Palette

Search for a command to run...

Scorpion Toxin Binder Design Dataset

Date

Organization

Anthropic

Scorpion toxin binder design is a scorpion toxin binder protein design dataset released by Anthropic in 2026. The related technical report is “De novo design of protein binders against scorpion toxins with Claude”, which mainly records scorpion toxin binder proteins designed de novo with Claude assistance and their experimental validation results.

The dataset contains 211 protein design samples, covering five scorpion toxins: Cn2, Ts1, AaH II, AaH I, and LqhIII. The research team, together with Adaptyv Bio and Twist Bioscience, evaluated the binding of proteins to target toxins using surface plasmon resonance (SPR) technology, and ultimately screened out 17 protein designs that met the binding criteria: 7 targeting Cn2, 4 targeting AaH II, 4 targeting AaH I, and 2 targeting LqhIII; no Ts1 binders meeting the criteria were screened out.

In addition to protein sequences and binding experiment results, the dataset also provides computational prediction scores, antibody competition experiments, polyreactivity ELISA assay results, sensorgrams, protein structure models, and supporting tables. It can be used for research on de novo protein design methods, binder screening, and correlation analysis between computational predictions and experimental results.

Dataset Composition

The dataset mainly contains the following files and directories:

  • design.csv: contains sequences, binding experiment results, and related parameters for 211 protein design samples;
  • in_silico_scores.csv: records the computational prediction scores corresponding to each design; - Cn2_competition_Twist_Bioscience.csv: records the antibody competition experiment results against Cn2;
  • BVP_ELISA.csv: records the polyreactivity ELISA assay results of the protein designs;
  • toxins.csv: records the sequences, categories, and design target information of the five scorpion toxins;
  • controls.csv: records information on control antibodies used in the experiments;
  • sensorgrams/: stores sensor experiment images, raw curves, kinetic fitting results, and some follow-up experimental data;
  • structures/: stores protein design models and predicted structures;
  • supporting_tables/: contains 12 supporting tables, recording information such as design constructs, protein expression, follow-up experiments, and related scores, interfaces, and thresholds.

Build AI with AI

From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.

AI Co-coding
Ready-to-use GPUs
Best Pricing

HyperAI Newsletters

Subscribe to our latest updates
We will deliver the latest updates of the week to your inbox at nine o'clock every Monday morning
Powered by MailChimp